As a game developer, I'm accustomed to building games around game mechanics. It's often a struggle to convey the importance of strong core game mechanics when they are not visible like beautiful art or animations. This article makes it clear that adding game mechanics as a (problem solving!?) wrapper to a non-game is akin to adding (problem solving!?) graphics to a game.
Nevertheless, I think that parallel can extend further in ways that might soften the article's point (just as the author did at the end). Some game design components, like clear state feedback and reward, are not so much game mechanics as they are signals of the mechanic's state (usually done with art in its general sense). In the same way, those signals could be incorporated into non-games to reflect the internal workings of their state.
So, in the end, I agree with the article and think it provides an insightful lens through which to view the parts of games relative to more tradition software.